Luc Bologna

Same issue: No luminance. If I don't mistake, the ability to change sensibility is missing too.

There is nothing insurmountable, I believe. Classes working with a tensor and returning a tensor should have a common abstract class inheriting from Layer. It would virtually and naively implement...

I fully agree your first point. From user side, everything should work with tensors as input or ouput. Tensorflow.Net need to take it into account internally, from Tensorflow.Net developpers pov....

I have a model that have one Tensor input, but two tensor output. Something like: var inputs = keras.Input(7, dtype: TF_DataType.TF_FLOAT); var outputs_pred = tf.keras.layers.Dense(3, activation: "relu").Apply(inputs); var outputs_pred2 =...

The separation into TensorFlow.Net and another package like TensorFlow.py, for example, is a good approach. I admit to being quite confused by the naming conventions and certain workarounds that need...

Hello :) Is there a way to help to maintain the repos ? BR,

Same issue, same solution. Thanks a lot @zzksdu :) PS: Prehaps an update of requirement.txt should be do ^^'